The real meaning of transcriptional silencing occurring at the rDNA locus has not been fully understood yet; in fact Pol II-transcribed genes, inserted into the rDNA, lie in regions where the DNA is heavily transcribed by Pol I; this excludes a simple silencing model based on transcription factor inaccessibility. We hypothesized that rDNA silencing represents the natural way to repress the Pol II-dependent ncRNA synthesis that could have potentially dangerous effects; in fact, it has been observed that ncRNAs, produced from the NTS region, are associated with altered epigenetic modifications of histones (19
) and induction of hyperrecombination in the ribosomal units (10
In order to clarify the physiological meaning of ribosomal silencing, we measured ncRNA production by Pol II in yeast strains differing in RNA polymerase I transcription rates. In addition, we wanted to better characterize the relationship between the two RNA polymerases' activity and reciprocal silencing. As reported in , overexpression of ncRNAs from C-PRO is observed for mutants defective in Pol I transcription, especially in the PSW mutant, where UAF is lacking (in this strain the rDNA copies are transcribed by Pol II). Actually, as reported with in vitro
and in vivo
), UAF binds DNA just on the region corresponding to the C-PRO promoter and this may be responsible for inhibition of Pol II transcription from this region (). This inhibitory role, performed by UAF, could be also exerted when the chromatin organization is completely lost (), during repression of H4 synthesis. The C-PRO-derived ncRNAs did not modify their synthesis, while E-PRO became active in the absence of nucleosomes. The capability of UAF to repress ncRNA synthesis may be relevant to the PSW phenotype. We hypothesize that the absence of UAF, necessary for the PSW phenotype, triggers ncRNA production that in turn induces rDNA recombination and copy number increase (10
). This essential step compensates for the low level of 35S RNA synthesis by Pol II in the PSW mutant, where Pol I is switched off (22
Chromatin modifications have been hypothesized to occur during rDNA silencing (19
); thus, we have evaluated whether alteration of histone acetylation occurs in mutants differing in 35S RNA transcription. When Pol II activity rises, producing more ncRNAs (strain PSW), a significant increase of histone acetylation is observed, specifically at the Pol II promoters (E-PRO and C-PRO) (); conversely, when Pol I transcription is maximal (strain ATU), the acetylation at the E- and C-PRO promoters goes down while it increases at the coding region (). Other regions engaged in neither Pol I nor Pol II transcription do not change their acetylation state significantly in the various mutants (ENH and NUC regions, all strains, ). These observations suggest that histone acetylation correlates with Pol II transcription at E- and C-PRO promoters when silencing is lost. The acetylation level of the coding region seems to directly depend only on Pol I efficiency. We also observed an effective loss of histone proteins when transcription of Pol I is high, while during Pol II activity this does not occur ( and ). Accordingly, it has been recently demonstrated that actively transcribed rRNA genes are largely devoid of histone molecules, and they are also associated with the high-mobility group protein Hmo1 (20
); moreover, significant chromatin reorganization has also been reported for active ribosomal genes (7
). In summary, our comparison between Pol I and Pol II transcription of ribosomal genes and ncRNAs shows that (i) Pol I, with its extremely high activity, works in a chromatin context where nucleosomes are lost or significantly acetylated and (ii) Pol II, whose efficiency of transcription at E- and C-PRO promoters is moderate, requires a significant increase of histone acetylation, without complete chromatin disruption, as shown in . This observation is also confirmed by the chromatin dissolution experiment () showing that in the strain where the H4 gene is shut down, only genes served by Pol II become more highly transcribed, while Pol I does not take advantage of an open chromatin condition because it is intrinsically able to remove nucleosomes. Previous reports (14
) have shown conflicts regarding the extent of Pol I transcription after chromatin disruption. Our results agree with those of Kim et al. (14
), even though different treatments, conditions, and measuring methods should be taken into account when differences are observed.
Finally, we investigated the role of DNA topoisomerase I in Pol I transcription and ribosomal silencing. We have demonstrated that in vivo DNA topoisomerase I activity correlates with Pol I transcription, suggesting that the swivel function goes together with the transcription machinery. Considering that top1 mutants show loss of transcriptional silencing, we can speculate that the swivel activity of DNA topoisomerase I is necessary to sustain Pol I transcription and, thus, that it is detrimental for Pol II activity.
In conclusion, our observations demonstrate that transcriptional silencing occurring at the rDNA locus represses ncRNAs with a strength inversely proportional to Pol I transcription. In addition, localized regions of histone hyperacetylation occur in E- and C-PRO elements when Pol II is active and in the coding region when Pol I is fully functional. The DNA topoisomerase I site-specific activity also follows RNA polymerase I transcription. We propose that repression of ncRNAs in response to active RNA polymerase I transcription represents a physiological regulative circuit ().
Relationship between Pol I transcription and ncRNA production. Schematic representation of the relationship between events in transcription by RNA polymerases I and II and histone acetylation in regulative and coding regions of rDNA in S. cerevisiae.
Previous studies have addressed the issue of ribosomal silencing by the analysis of reporter genes artificially inserted in the ribosomal locus (2
). The experiments reported here indicate that 35S RNA transcription by RNA polymerase I represents the natural control of ncRNA production, associated with genome instability and possibly aging in yeast. In addition we identified a new epigenetic marker of Pol II activity, H3 and H4 histone acetylation, within the NTS region at the ncRNA cryptic promoters.